Asian Research Thesis Index

Abstract

A Business Process (BP) is a set of logically related activities performed to achieve a well-defined business outcome. The activities in a BP have control-flow and data-flow dependencies which are defined in the workflow model of the BP. In the service-oriented environment, these activities are performed by invoking Web services. Typically, BP development involves working with high-level process specifications, and manually elaborating these specifications into an implementation level workflow design. After this, each activity in the BP workflow is binded to the appropriate Web service operations for executable BP code generation. This is a very elaborate process requiring substantial manual effort and investment in the IT infrastructure, personnel and resources. This is specifically challenging for small and medium enterprises which often do not have adequate resources for coding, administration, and hosting of their BPs. Cloud infrastructure and serviceoriented middleware can be leveraged for rapid and cost effective development of BPs for such organizations. Our objective in this dissertation is to develop an integrated framework for collaborative BP development in a services cloud environment. The proposed framework enables an organization to compose, deploy, and execute their BPs using the Web services and resources available in the cloud. For collaborative BP development, the proposed framework makes use of the knowledge of existing BPs in a given domain to learn common execution patterns of different types of BPs and uses the learned patterns for composing a given type of BP for an organization based on its requirements. The proposed framework aims to assist organizations by enabling automated composition of BPs. Given the requirements and constraints of the user organization and the available Web services, the proposed framework helps the user organization to develop an implementation level BP workflow, which includes the required set of activities, the control-flow and data-flow between these activities, as well as mapping of the activities to appropriate Web services. This implementation level workflow can be translated directly into an executable BP for deployment and execution. The underlying approach for collaborative BP development is designed for structured BPs as well as knowledge-driven BPs. Structured BPs have a well-defined workflow which is predictable and stable over time. In contrast, knowledge-driven BPs are knowledge and data centric and their workflow structure and execution may evolve dynamically based on the changes in the environmental context, user requirements and composition goals. To support evolution of a knowledgedriven BP, the proposed framework also supports on-the-fly process composition to extend the BP workflow in a situation and context aware manner.

Add/Update Thesis